IOP Conference Series: Earth and Environmental Science

PAPER • OPEN ACCESS

3D Numerical Analysis for Rock Support Design of an Underground

Powerhouse Cavern in Complex Geological Conditions
To cite this article: E Yildiz et al 2021 IOP Conf. Ser.: Earth Environ. Sci. 833 012091

View the article online for updates and enhancements.

This content was downloaded from IP address 109.201.133.100 on 07/09/2021 at 12:33

Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

3D Numerical Analysis for Rock Support Design of an

Underground Powerhouse Cavern in Complex Geological
Conditions

E Yildiz1, K Thermann1 and I Kamuf2

1
Tractebel Engineering GmbH, 61118 Bad Vilbel, Germany
2
EnBW Energie Baden-Württemberg AG, 70657 Stuttgart, Germany

[email protected]

Abstract. A new underground powerhouse that combines a run-of river scheme with a pumped
storage scheme is planned in course of the Renewal of Forbach HPP in Germany. The
powerhouse cavern is designed to have a maximum span and height of 18.9 and 42 m, and a total
length of 123 m. The rock mass conditions were found favourable in general. However,
comprehensive investigations revealed weakness planes crossing the cavern. The support system
was determined by joint evaluation of empirical methods, wedge analysis and the 2-D finite
element analyses. Due to the complicated geometry of the cavern and the orientation of the
weakness planes, the effect of weakness zones on the stability and deformations could not be
evaluated by only 2-D analyses. Implementing an additional 3-D analysis and comparison of the
results with the common 2-D approach allowed a more realistic assessment of the rock
behaviour, especially under the existence of the weakness zones through the cavern. A very
detailed 3-D finite element model for the whole cavern is still not practical, so a relatively larger
mesh was adopted. Nevertheless, a reduced model for a limited part of the cavern was also
analysed by using a finer mesh. Obtained results were similar indicating the suitability of the
adopted numerical models for the rock support assessment. The deformations and the size of the
plastic zones were identified as significantly larger for weakness zones passing through the
cavern. For these intersection areas additional rock support measures including pre-stressed
anchors were therefore planned.

1. Project Information
The Rudolf-Fettweis Plant by Energie Baden-Württemberg AG (EnBW) in Forbach has great potential
to increase storage capacity and thus contribute to the energy revolution. The run-of-river, storage, and
pumped storage power plant consists of four individual power plants built between 1914 and 1926. The
two major units will be replaced by the new power plant.
In course of the renewal and extension, a new underground powerhouse is planned. The new
powerhouse cavern will have a length of 123 m, maximum span of 18.9 m and a maximum height of 42
m. The cavern will be located approximately 325 m below ground in crystalline rocks of the Forbach
Granite. The expansion of the equalising reservoir in Forbach to include a cavern storage facility leads
to an additional storage volume and thus a significant increase in possible turbine operation. The project
is currently in the approval process. In case of approval and a positive investment decision by EnBW,
construction could start in 2022/2023 at the earliest. The general layout of the project is shown in
figure 1.

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

Figure 1. General Layout of the Project.

2. Geological Information

2.1. Geotechnical Investigations

There have been several investigation campaigns performed between 2010 and 2019. 32 boreholes were
drilled in total with depths up to 400 m. The boreholes were utilised to perform in-situ testing like
borehole scanning, Lugeon-tests, dilatometer tests, hydraulic tests (slug, pulse, pump tests) and in-situ
stress measurements. Furthermore, seismic reflexion and seismic refraction surveys were performed
from the surface going along with a detailed geological mapping of the project area. The field
investigations were accompanied by a comprehensive laboratory program comprising rock index tests,
UCS and triaxial tests, Brazilian-tests, tests regarding abrasiveness, freezing-thawing and finally
petrographic analyses on thin sections.

2.2. Geology of the Site

The project area is located in the northern part of the Black Forrest which is a low mountain rage formed
by sedimentary rocks of the Upper Permian and Lower Triassic rocks. These formations are underlain
by granitic bedrock, the so-called Forbach-Granite which has a typical reddish appearance. Aplite dikes
can be found in the bedrock as well as several weakness zones caused by weathering processes which
locally alter the bedrock significantly.

Figure 2. Geological Profile and Cross-Section of the Powerhouse Area.

2
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

All underground structures are located completely within the bedrock of Forbach-Granite. Figure 2
shows the geological profile and cross section of the powerhouse cavern with borehole logs and
indication of several weakness zones (dotted lines). As can be seen, in total 3 weakness zones are
expected in the cavern area, 2 running directly through the cavern excavation and one located just in
front of the SW-end wall of the cavern. These weakness zones are supposed to have been formed along
some major joints. The strike direction of the weakness zones is almost perpendicular to the cavern axis.
Thus, the cavern is orientated very favourable to these zones reducing the potential impact on excavation
and support measures to the best possible extent.

3. Rock Mass Parameters

According to the obtained data from the comprehensive geological and geotechnical investigations, the
rock mass conditions around the underground powerhouse cavern have been evaluated as favourable in
general. The cavern excavations will be in general within fresh and relatively unweathered granite rock
mass, which is referred as G1, with an estimated average GSI rating of 60. Based on the results of the
investigations and geological evaluation, three weakness layers with a higher degree of weathering have
been identified in the cavern area. These weakness layers are estimated to have a thickness between 0.6
– 1.5 m. The granite rock mass in the weakness layers is referred as G4 and as highly weathered and
partly disintegrated. The GSI rating has been reduced to 43 for the weakness zones.
Due to the orientation of the powerhouse adopted in design, these weathered layers, identified by
core logging and surface mapping, were primarily oriented within a joint set which is perpendicular to
the axis of the cavern and cuts the cavern at a favourable angle. Investigations have also revealed that
larger areas of weathering, decomposition or shear zones, which may be more critical in terms of
stability and deformations, have not been expected.
For the assessment on the stiffness of the rock mass classes, a total of 11 borehole dilatometer tests
have been carried out in 6 different boreholes. Empirical estimates have also been made according to
Hoek et al. (2002) and Hoek & Diederichs (2006). Obtained results from different methods were in
agreement from an engineering point of view, as summarized in table 1. The deformation modulus has
been defined as 17.5 GPa and 0.9 GPa for G1 and G4 rock masses based on an overall evaluation.

Table 1. Deformation Modulus for Rock Mass (in GPa).

Granite - G1 Granite – G4
(weakness zone)
GSI = 60, UCS = 84 MPa GSI = 43, UCS = 10 MPa
Hoek et al. (2002) 16.3 2.1
Hoek & Diederichs (2006) - Simplified 20.4 5.2
Hoek & Diederichs (2006) – with
16.6 0.9
Ei = 32 GPa (G1), 4.5 GPa (G4)
Hoek & Diederichs (2006) – with
18.6 0.8
modulus ratio, MR = 425
In-Situ Test Results 17.3 0.67
Selected for Analysis 17.5 0.9

In order to estimate the shear strength of the rock mass, Hoek-Brown criterion based on GSI has been
adopted. As the Hoek-Brown failure law is non-linear, it has been preferred to use linear Mohr-Coulomb
failure criterion, whose parameters are commonly derived for a certain stress range as an approximation.
The stress conditions corresponding to the overburden of 325 m have been considered in order to
determine the equivalent Mohr-Coulomb parameters to represent peak shear strength. For fresh to
slightly weathered granite, G1, the residual shear strength has also been considered and the residual
parameters have been determined according to Crowder & Bawden (2004). Adopted shear strength
parameters in the calculations are given in table 2 below.

3
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

Table 2. Shear Strength Parameters for the Rock Mass.

Granite - G1 Granite - G4
Peak Residual Peak Residual
Cohesion, c (MPa) 2.3 1.5 0.9 0.9
Friction Angle, φ (deg.) 56 46 35 35

4. Rock Support Design

4.1. General Design Approach

In the current engineering practice, there are different methods regarding the evaluation of support
requirements for the underground excavations in rock. In the design, an overall assessment has been
made considering basically a 3-fold approach. The rock bolt lengths and patterns have been determined
(i) using empirical methods based on rock mass classification, (ii) wedge-type failures controlled by
main discontinuities and (iii) the extent of plastic zones around the excavation calculated by the finite
element analyses. The empirical methods and wedge analyses are not covered in this paper due to lack
of space, but main focus is kept on the 2-D and 3-D finite element analyses, as discussed in the rest of
the paper.

4.2. 2-D Finite Element Analyses

The numerical modelling of underground structures is commonly performed in two dimensions
considering plane-strain conditions as a simple but reliable approach. The finite element models have
been established for two critical cross-sections as the pumped storage and run-of-river sections which
will provide insight to the overall behaviour of the rock mass in real conditions. The rock mass
parameters for G1 rock mass, which is dominant at the cavern location, have been considered in the 2-
D calculations.
Within the scope of the investigations, in-situ stress measurements have been carried out using the
borehole relaxation method using triaxial cells. According to the obtained results, the maximum
principal stress is oriented horizontally in NW-SE (~ 140°), which is almost perpendicular to the
longitudinal axis of the cavern. The minimum principal stress is also oriented almost horizontally
(approx. 10°) in the NE-SW direction (~ 50°). Considering the depth of 325 m for the cavern, the
following stress values have been adopted in the analyses: Vertical Stress (v) = 8.27 MPa, Horizontal
Stress along the cross section of the cavern (1h) = 11.32 MPa, Horizontal Stress in the longitudinal axis
of the cavern the cavern (3h) = 3.75 MPa. Considering the large depth of the cavern and high stresses,
field stress option with the mentioned values have been used in the numerical analysis, by also applying
fixed boundary conditions around the numerical model as required for a constant stress analysis. The
staging of the cavern excavation and the sequence of excavation – support installation have also been
considered in the analysis in order to make a realistic assessment on the forces on the support elements.
An elastoplastic material model has been used to represent the rock mass behaviour and the shear
strength of the material has been defined considering both peak and residual values based on the
respective parameters as defined in Section 3. The stress conditions, deformations and plasticised
(yielding) zones around the cavern at two representative cross-sections have been examined under these
assumptions. The deformations and the yielded elements after the excavation are shown in figure 3 for
the pumped storage and run-of-river sections.
It can be seen from the results, for the pumped storage section which is critical due to its excavation
height, the rock mass yielding in the roof and the side walls extends up to 3.5 m and 2.6 m respectively.
The largest total displacements at the excavation boundary occur around the middle of the side walls
with a maximum of 20 mm. The total displacements in the roof are relatively smaller as 8-10 mm. In
the wedge analysis which has been separately performed, sliding wedges with a depth of up to 9 m have
been taken into account in the cavern roof (when scaled according to persistence). It has been observed
that the plasticization of the rock mass only extends up to 3.5 m which will not lead to formation of new

4
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

joints due to yielding. In this regard, larger wedges are not expected to be formed than those that have
been already considered, based on the wedge analysis results.

(a) (b)

Figure 3. Total Displacements and Yielded Elements After Excavation at a) Pumped Storage
Section, b) Run-of-River Section.

The maximum displacements of the run-of-river section are 6-8 mm in the roof and 11 mm in the
side walls. The yielded zone extends up to 2.4 m within the rock mass at the roof, while the plasticization
in the side walls extends up to 2.0 m. Due to the lower height of the side walls, the displacements and
the extent of the yielding zone are significantly smaller compared to the pumped storage section. For
the roof, on the other hand, the displacements and the depth of the plastic zone are similar to those for
the pumped storage section.
The plane-strain finite element analyses showed that the powerhouse cavern can be considered as
stable and self-supporting, when the rock mass is considered as a continuum. It can be said that a support
system does not seem necessary under the assumptions of the 2-D finite element analysis. In this case,
it has been found reasonable and safe to determine the rock support measures primarily based on the
wedge analyses where the discontinuities have been taken into account.
It has been decided that the rock bolts should reach at least 1 m beyond the critical wedge or the
plastic zone to be effective on the long term. The pattern and length of the rock bolts have been
determined as:
- For the roof: 32 mm diameter, L = 8 m, Pattern of 1 bolt per 2.25 m2
- For the side walls: 32 mm diameter, L = 6 - 8 m, Pattern of 1 bolt per 5 m2
- For the end walls: 32 mm diameter, L = 6 m, Pattern of 1 bolt per 5 m2
A shotcrete layer of 20 cm with double-layer mesh reinforcement has been found necessary for the
roof section only, which will provide resistance to smaller rock wedges and improve the rock bolt
protection at the roof. For the walls, no shotcrete protection has been considered as necessary for
favourable rock conditions of granite. However, shotcrete of 10-15 cm has been suggested for potential
weak areas or weathered zones, which are expected to be locally encountered and limited to small
extents.
The loads acting on the suggested support system have been examined with an additional analysis
where the rock bolts and the shotcrete elements have been included in the numerical model. Figure 4
shows the results of deformations and yielded zones in the case of support application.
Results indicate that the applied rock support measures have hardly any influence on the
deformations. On the other hand, there has been a decrease of about 1 m in the extent of the yielding
zones for the maximum cross-section. Support elements have been checked against overstressing due to
the excavation process. Both the rock bolt and shotcrete elements have been loaded under their capacities
indicating that the suggested support system is suitable and adequate.

5
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

Rock Bolt

Rock Bolt

Shotcrete 20 cm

Shotcrete 20 cm

Rock Bolt

Rock Bolt

(a) (b)

Figure 4. Results for Supported Case at a) Pumped Storage Section, b) Run-of-River Section.

5. 3-D Finite Element Analyses

As discussed in Sections 2 and 3, although the rock mass conditions are favourable around the cavern
in general, there are three weakness planes identified, two of which are cutting through the cavern
excavation. The strike direction of these planes are perpendicular to the cavern axis and the dip angle
makes an angle of about 50° with the horizontal, as indicated on figures 2 and 5. The performed studies
explained previously do not consider these weakness layers. For a complete study on the rock mass
behaviour during the cavern excavations, the effect of these weakness zones with lower stiffness and
strength need to be assessed.
The weakness zones have been identified as continuous in their own longitudinal direction. However,
a two-dimensional approach for this purpose is not convenient due to the relatively small width of these
weak layers and their orientation as well. Consideration of only the intersection of these weakness planes
for the cross-section in a numerical model may lead to underestimation of the adverse effect of these
planes. Consideration of a larger area for the weakness zones in a single cross section may on the other
hand result in over-conservative results. Modelling in 2-D longitudinal section is also not an alternative
as it cannot capture the realistic behaviour around excavation. In this regard, a three-dimensional finite
element analysis has also been performed for the cavern excavation. The model information and the
obtained results are given in the following sections.

5.1. Numerical Model

The finite element model has been established considering the cavern geometry and the connecting
galleries as realistic as possible. The weakness planes have been modelled as solid elements, with a
thickness of 2 m at the designated locations. The main goal of this analysis was to examine the effect of
the weakness zones on the stresses, deformations and yielding zones, so the numerical model did not
include the support elements, in order to simplify the modelling and staging in the analysis.
Similar to the 2-D finite element model, the 3-D numerical model had also fixed boundaries around
and the initial stress conditions have been applied as constant field stresses in three different directions
with the values and directions as mentioned in Section 4. A very detailed numerical model for the whole
cavern geometry, using a mesh size similar to the one adopted in 2-D analyses is not practical even with
current computers. In this regard, a relatively large mesh has been adopted for the numerical model. The
distance between adjacent nodes was between 1.3 – 4 m at the excavation boundary. Besides, a reduced
model for a limited portion of the cavern including the critical cross-section has also been established
by using a finer mesh size, in order to ensure that the adopted finite element model has sufficient
accuracy, by comparing the results. The maximum distance between the adjacent nodes was reduced to
1.6 m at the excavation boundary. The finite element models are shown in figure 5, where the weakness
zones are also indicated in the profile.

6
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

(a) (b)

(d)

(c)

Figure 5. Finite Element Model and Weakness Zones in Profile. a,b)The whole numerical model;
c) Profile View of Weakness Layers; d) Reduced numerical model with finer mesh size

5.2. Results of the Analysis

Obtained results of total displacements and the yielded elements are shown in figure 6. It can be observed
that the displacements significantly increase at the areas of weakness planes, both at the sides and roof
of the cavern. Similarly, the extent of yielded elements increases, especially around the cavern side walls
at the second (middle) weakness plane location. The first weakness plane, which is in front of the cavern
excavation has almost no effect, although it passes close to the base corner of the cavern.
Obtained results for the two numerical models have been found as very similar, indicating the
suitability of the adopted numerical model for an accurate assessment of the three dimensional
behaviour. The size of the finite element mesh has been kept at an optimized level, which has allowed
to capture the behavior at the weakness zones and crossing tunnels in an elaborative way, with also
reasonable analysis run time.
The maximum displacements at the sides and the roof are 77 mm and 37 mm respectively, which are
significantly larger than the ones obtained in 2-D analysis as 20 mm and 10 mm. However, this drastic
increase of displacements and yielding zones is only at the intersection of the weakness planes. The
maximum extent of the yielding zones is about 10 m at the side wall and 6 m in the roof at the intersection
area. These values are considerably higher than those by 2-D analysis as 2.6 m and 3.5 m. Similar to the
displacements, the enlargement in the plastic zone is limited to the immediate vicinity of the defined
weakness layers. The results for the rest of the 3-D numerical model are similar to the ones by 2-D
analysis.
In the light of these results, it can be concluded that the weakness planes are not expected to adversely
affect the overall rock mass behaviour. The effect of locally adverse zones which may be encountered
during the excavations is expected to similarly have a limited impact in small extent only. Increasing
the pattern of the envisaged rock bolts and/or shotcrete thickness, or, if required, the additional
application of pre-stressed anchors have been considered as potential mitigation measures for weakness
zones or zones of adverse rock conditions. Double corrosion protected thread-bar anchors, with a steel
grade of 555/700 MPa, diameter of 63.5 mm and a nominal load of 1 MN, or alternative cable anchors
have been suggested as additional support for the intersection area with weakness planes. The anchor
lengths have been suggested as 15 m and 20 m for the roof and side walls respectively, with a bonded
length of 5 m.

7
Mechanics and Rock Engineering, from Theory to Practice IOP Publishing
IOP Conf. Series: Earth and Environmental Science 833 (2021) 012091 doi:10.1088/1755-1315/833/1/012091

In the areas of the connecting galleries or enlargement sections, there is no considerable increase in
the deformations or the yielded elements. In this regard, due to the large dimensions of the cavern
excavation, the critical areas are the roof and the sidewalls of the pumped storage section. This also
indicates that the 2-D results can be used as a reliable tool in case of rather homogeneous conditions of
rock, without any weakness zones or planes for similar geometries.

Total Displacements
≈30 Yielded Elements

≈37 ≈52

≈77 ≈6 m
(vertical)

≈10 m

(a)
Total Displacements
≈37 Yielded Elements
- Displacements
are in mm.

- Gray zones ≈86

indicate yielding
elements.

(b)

Figure 6. Total Displacements and Yielded Elements after Excavation. a) Model for the whole
cavern geometry; b) Reduced Partial Model for Critical Area

6. Conclusions
The rock support design of an underground powerhouse cavern has been carried out by conventional
methods and additional three-dimensional finite element analyses. A 3-D numerical analysis was
considered as mandatory in order to assess the impact of weakness planes in the powerhouse area,
revealed by the geotechnical investigations. The underground cavern is located in granite rock mass
with favourable conditions in general. The results by common 2-D approach have indicated small
displacements and local yielding zones due to the good quality rock mass. The suggested support system
included the common shotcrete lining and rock bolts around the excavation. Performed 3-D finite
element analysis has allowed to examine the impact of the weakness layers on the stability and
deformations. Obtained results have shown that two weakness planes cutting through the cavern have a
considerable impact on the cavern excavation in terms of increased deformations and extent of yielding
zones. As mitigation measures, additional support application potentially including also pre-stressed
anchors is suggested for these areas. Results have also indicated that the weakness planes are not
expected to have an adverse effect on the overall rock mass behaviour and stability.

References
[1] Hoek E, Carranza-Torres C and Corkum B 2002 Hoek-Brown Failure Criterion - 2002 edition
Proc. NARMS-TAC Conference Toronto 267-273
[2] Hoek E and Diederichs M S 2006 Empirical Estimation of Rock Mass Modulus Int. J. Rock Mech.
Min. Sci 43 203-215
[3] Crowder J J and Bawden W F 2004 Review of Post-Peak Parameters and Behaviour of Rock
Masses: Current Trends and Research